In general, there is a lack of efficient, high-power lasers in the 1100-1500 nm region with a controllable linewidth. Lasers in the 1100-1500 nm spectral region are difficult to obtain, since many materials don't lase in this region and those that do lase in parts of the spectral region such as bismuth co-doped silica or Yb-doped silica, do so inefficiently. One way of obtaining photons in this spectral region is through the nonlinear process of Stimulated Raman Scattering which acts to shift the initial pump wavelength out to longer wavelengths. This process, which occurs at high optical intensities, involves the coupling of light propagating through the nonlinear medium to the vibrational modes of the medium. The result is re-radiated light which is shifted to a different wavelength. Light upshifted in wavelength is commonly referred to as a Stokes line, whereas light downshifted in wavelength is referred to as an anti-Stokes line. To-date, a controllable linewidth, high-power Raman laser with output powers approaching 100 W has not been reported.
The typical Raman amplifier tends to be in several forms. The first form is seeded with an initial pump signal of relatively low power (zeroth order Stokes line) that is free-space coupled or spliced into the system. Multiple orders of Stokes lines are then created in one or more Raman resonators. Each Raman resonator in the system is defined by a pair of Bragg gratings centered at the wavelength of the Stokes order involved. The output of one Raman resonator is injected into the Raman resonator centered at the next highest Stokes order. The highest order Stokes line generated is the output of the laser. Such amplifiers typically tend to have low Raman conversion efficiencies and low output powers (<5 W) due to the relatively low intensity of injected pump signal in the core. Raman fibers hundreds of meters in length are often required to enable adequate conversion of the pump signal to a longer wavelength. Such lasers also tend to have broad linewidths, since the Raman process is initiated by broad-band spontaneous Raman scattering within the fiber. It has also been observed that, as the power in the system is increased and/or higher-order Stokes lines are generated, the linewidth of the output tends to further broaden. To conclude, this type of Raman laser tends to have low output powers in addition to linewidths, which are not controlled and are broad.
A variant of the above are all-fiber Raman systems where the pump is either generated or amplified by a rare-earth amplifier that is spliced directly onto the Raman resonators (Nicholson, US Patent Application No. 2010/0284060A1 and Nicholson, et al., Optics Letters 35(18)(2010) 3069). This type of system has the potential of generating high output powers (>50 W) because of direct injection of high power levels of the zeroth order Stokes directly into the core of the fiber via an amplifier. Highest power levels were achieved by Nicholson et al., who demonstrated 81 W of output power at 1480 nm when 162 W of 1117 nm pump from an amplifier was injected into 120 m of Raman filter fiber. Although high power levels were achieved, a long length of Raman fiber was still required in order to enable sufficient buildup in the Raman cavities of the various Stokes lines from spontaneous Raman scattering. The resultant linewidth of these lasers is broad, since no measures are taken to control linewidth broadening. To conclude, this kind of laser is capable of high output powers but the linewidths tend to be broad.
Another variant of a Raman amplifier involves both seeding with the desired output signal (Nth order Stokes) and a pump signal (zeroth order Stokes) through either a wavelength division multiplexer (WDM) or an optical circulator (see for example, Taylor et al., US Patent Application No. 2011/0038035A1). In such systems, one or more stages of Raman amplification may be necessary in order to generate the N−1st order Stokes signal necessary for amplification of the Nst order Stokes seed. Because both WDMs and optical circulators are power limited, the amount of pump signal (zeroth order Stokes) and desired output signal (Nth order Stokes) that can be fed into the system is limited. Because of low levels of pump signal in the system, output power levels are limited, efficiencies tend to be low, and extremely long Raman fibers (100 m or more) are necessary. Output power levels are also limited by Stimulated Brillouin Scattering for narrow linewidth signals due to the long Raman fiber. Relative to an unseeded system, the linewidth of the amplified output signal is controllable to a certain degree, since the seed signal will dominate the spontaneous Raman scattering. Even so, linewidth broadening will still occur because of four-wave mixing. To conclude, this sort of system is capable of lower output powers with some control of the linewidth.
Another variant is to seed a system with power from a rare-earth-doped oscillator that is spliced directly onto a Raman resonator (Mead, US Patent Application No. 2011/0122482). In this patent application, the main focus is on generating closely spaced wavelengths from multiple Raman fiber amplifiers through stretching of fiber to enable spectral beam combination of eye-safe lasers. One embodiment is shown where the system is seeded with the desired output wavelength (Nth order Stokes) and is Q-switched to generate pulsed light. No continuous wave configuration is discussed. The rare-earth-doped oscillator, which is unseeded, is co- and counter-pumped with diodes. The output wavelength of the oscillator is determined by the Bragg gratings that form the cavity and will have a broad linewidth, since it is seeded with amplified spontaneous emission. The rare-earth-doped oscillator/Raman resonator described in this embodiment should be capable of high output power because of direct injection of the zeroth order Stokes from the rare-earth-doped oscillator into the core of the fiber in the Raman resonators but, once again, the Nth order Stokes seed will experience linewidth broadening as it is amplified since no measures are taken to mitigate four-wave mixing. This configuration will also be power limited relative to what would be achievable using a comparable rare-earth-doped amplifier because of thermal issues associated with power buildup in the cavity of the rare-earth-doped oscillator. Also, for high-power applications, fully nested Raman cavities as they appear in this patent application will experience higher thermal stresses than a less overlapped configuration. In addition, the embodiment in this patent application may experience damage upstream, since no measures are taken to mitigate light leaking out of the gratings and propagating backwards. To conclude, the system described will be capable of high-power output pulses (but not continuous wave operation), but because no measures are taken to control four-wave mixing, broadening of the signal linewidth will occur. Also, high power levels within the rare-earth-doped oscillator and the Raman cavities may be a limiting issue.
An important narrow-linewidth application for Raman lasers is the generation of 1178 nm for guide star (e.g., sodium guidestar) laser applications. This is important since the resolution of terrestrial telescopes is limited by wave front distortion caused by atmospheric turbulence. This distortion can largely be overcome by the use of adaptive optics in which the surface of a deformable telescope mirror is varied as a function of time to compensate for atmospheric turbulence through which light from distant objects must travel. Measuring the distortion requires that there be a bright optical source in the sky, such as a bright star, located close to the object to be observed. Since bright stars are infrequently located close to objects of interest. An alternative is to energize a layer of sodium atoms which is naturally present in the mesosphere at an altitude of around 90 kilometers. The sodium atoms then re-emit the laser light, producing a glowing artificial star whose radiation can provide a wavefront reference to enable correction of the image for atmospheric induced aberrations.
The sodium guidestar laser application is very challenging in that output powers on the order of 50 W of 589.15908 nm on the sodium D2a line with a linewidth of 10 MHz is required. For two-line systems, an additional 10 W of 589.15709 nm on the sodium D2B line with a linewidth of 10 MHz is desired. One very successful method for generating 50 W of 10 MHz linewidth 589 nm involves the use of traditional rod laser technology and sum-frequency generation via a nonlinear crystal. A state-of-the-art system developed at the Air Force Research Laboratory (Denman, et al., U.S. Pat. No. 7,035,297) contains 1064 and 1319 nm resonant cavities in addition to a doubly resonant sum-frequency generation cavity containing a lithium triborate crystal. To maintain lock on the D2a line of sodium, multiple Pound-Drever-Hall locking loops are utilized. A maximum of 50 W of 589 nm was achieved by this system. The major drawback associated with this system is its size and complexity.
The most successful attempt at addressing the requirement for the sodium guidestar laser using fiber was accomplished by the European Southern Observatory (ESO) (Feng, et al., Optics Express 17(21)(2009) 19021). In this concept, a Raman amplifier is directly core pumped with 1121 nm in a counter-pumped configuration while being seeded with narrow linewidth 1178 nm through a wavelength division multiplexer (WDM). The amount of narrow linewidth 1178 nm that can be generated was limited to less than 39 W. Coherent beam combination has been used to generate 26 W of 589 nm from two 1178 nm sources and greater than 50 W from three Raman fiber amplifiers. The linewidth of the 589 nm was found to be less than 2.3 MHz. The bottleneck associated with this technique is a power limitation associated with the WDM that clamps the amount of 1121 nm that can be injected into the system. This, coupled with problems from SBS resulting from a Raman fiber of length 150 meters, results in relatively low output powers of 1178 nm. As a result, in order to generate the levels of 589 nm desired by the various telescopes, coherent beam combination of multiple Raman amplifiers is necessary. The result is a system of increased complexity.
Since astronomical telescopes are located in remote sites and often operate under difficult conditions, it is desirable to use a compact, maintenance-free, and rugged laser for the guide star system, such as a fiber laser. Because there are no fiber gain media that lase directly at 589 nm, the most promising way, to date, to achieve this is through second harmonic generation of 1178 nm. The present invention has the potential to generate 100 W of narrow linewidth 1178 nm from an all-fiber Raman laser for frequency doubling to 589 nm on the D2a line.
Another exemplary application in the 1200-1300 nm region is remote sensing of the water content of vegetation on earth from space. The goal is to improve the understanding of the biophysical and ecological processes governing the linked exchanges of water, energy, carbon and trace gases between the terrestrial biosphere and the atmosphere by improving satellite data products for models. 1240 nm having linewidths on the order of 100 MHz to 1 GHz in conjunction with 858.5 nm is of interest since the optical index R858.5/R1240 is mainly driven by the water thickness with smaller effects due to cellulose, lignin, and protein variation. The wavelength 1240 nm has been generated in the past using several methods to include external cavity diamond Raman lasers, an intra-cavity Raman laser, GaInNAs semiconductor diode lasers, optically pumped GaInNAs/GaAs and a Cr:forsterite multi-terawatt amplifier laser. Output powers are typically less than 5-6 W with the linewidth being determined by properties of the resonating cavity. High power levels of 1240 nm can be achieved in an all-fiber system with a controllable linewidth by the present invention.
Another exemplary application in the 1300-1500 nm region is the expansion of telecommunications bandwidth into the O, E, and S bands. This will require multiple closely spaced lasers having linewidths well exceeding a GHz in this spectral region. At the present time, some work has been done with seeded Raman lasers at 1300 nm. Such lasers are inefficient and typically of low output power because of having to insert and remove light through power sensitive components. Unseeded Raman lasers using phosphosilicate fiber by itself and in conjunction with germanosilicate fiber have been used to generate light in the 1400-1500 nm region. Such lasers can be of rather high output power but are of broad linewidth because of the lack of a seed or other measures to control the linewidth. The invention in this application can be utilized to provide a series of lasers having controllable linewidths in the 1300-1500 nm region to enable expansion of the telecommunications industry into other bands.
The current invention is aimed at overcoming the shortcomings associated with the current state-of-the-art to obtain high power (>50 W) from a Raman amplifier while controlling linewidth broadening.